ReMood

🎧 ReMood: Emotion-Based Music Player 🎶

ReMood is a Python-based Machine Learning project that plays music dynamically based on the user's real-time facial emotions. It uses computer vision and emotion recognition to personalize the music listening experience, making it more engaging and emotionally aware.

🚀 Features

• Real-time facial expression recognition using a CNN model
• Emotion detection across 7 classes: `angry`, `disgust`, `fear`, `happy`, `neutral`, `sad`, `surprise`
• Personalized music playback based on detected emotions

🗃️ Dataset

• Source: FER-2013 Dataset
• Images: 48x48 pixel grayscale face images
• Total Images: ~36,000
• /train Data- 70%
• /test Data- 30%
• Each folder contains 7 emotion subfolders

🏗️ Model Architecture

Input: (48, 48, 1)
→ Conv2D (128) + ReLU + MaxPooling + Dropout
→ Conv2D (256) + ReLU + MaxPooling + Dropout
→ Conv2D (512) + ReLU + MaxPooling + Dropout
→ Conv2D (512) + ReLU + MaxPooling + Dropout
→ Flatten
→ Dense (512) + ReLU + Dropout
→ Dense (256) + ReLU + Dropout
→ Output Dense (7) + Softmax

📈 Results

• Best Validation Accuracy: ~65%
• Model saved with: .json structure + .h5 weights
• Confusion matrix used for performance evaluation

📜 License

This project is part of an academic mini project at Atharva College of Engineering, Mumbai. Feel free to use and adapt it for educational or non-commercial purposes.